By Catherine A. Cardno, Ph.D.
Transportation engineers at North Carolina State University in Raleigh have proposed a fourth light on traffic signals. This light would better streamline traffic at intersections, as human-driven cars increasingly need to interact with self-driving, autonomous vehicles.
“Autonomous vehicles are arriving, and they will be sharing our roadways with us. This concept aims at using AV technology to help with traffic control and make traffic operations more efficient, safer, and more environmentally friendly by reducing fuel consumption,” for AVs and human-driven vehicles, says Ali Hajbabaie, Ph.D., an associate professor in the department of civil, construction, and environmental engineering at N.C. State.
Hajbabaie was the project manager for the study and shared responsibility for its conceptualization, development, and modeling with Leila Hajibabai, Ph.D., an assistant professor in the department of industrial and systems engineering at the university.
The light would work on two levels. “The main purpose of the fourth light is to inform human drivers that a new traffic regime is in place,” Hajibabai says. But while the color of the light would cue humans to “platoon” with surrounding vehicles, the system would wirelessly interact with autonomous vehicles by way of a distributed computing framework — that is, using the computing capabilities of each AV approaching an intersection and the traffic signals in tandem.
“Utilizing our proposed distributed coordination framework, we could successfully find near-optimal signal-timing parameters,” says Ramin Niroumand, a Ph.D. candidate and research assistant in the university’s department of civil, construction, and environmental engineering. Niroumand was responsible for the study’s modeling, implementation, simulation, and analysis.
When present at an intersection, the AVs would be in constant communication with that intersection’s computer control system, helping each signal operate as efficiently as possible for the existing traffic situation.
When a certain proportion of vehicles approaching an intersection are AVs, they would coordinate their movements with one another through the intersection and trigger the use of the white light so that human drivers knew to follow the behavior of the vehicles in front of them.
“When enough AVs are present at an intersection, they can negotiate a safe right of way for groups of human-driven vehicles and navigate them through the intersection,” Hajbabaie explains.
The computer-based traffic simulator system has been tested at a small scale by the researchers, with their findings shared in the paper, “White Phase Intersection Control Through Distributed Coordination: A Mobile Controller Paradigm in a Mixed Traffic Stream,” which was published in February 2023 in the journal IEEE Transactions on Intelligent Transportation Systems. Niroumand was the lead author of the paper.
By platooning AVs and human-driven cars, stop-and-go traffic would be reduced around intersections, which would reduce fuel consumption, according to the researchers.
The traffic efficiencies would increase with the proportion of AVs traveling through an intersection. For example, the study’s simulations found that when 10% of vehicles were autonomous, delays were reduced by 3%. When 30% of vehicles were autonomous in the tests, delays were reduced by 10.7%. “Our simulation results show that the concept improves traffic operations,” Hajbabaie explains. “These improvements are not specific to AVs.”
The researchers believe that abandoning the existing visually oriented traffic light infrastructure will not necessarily be an option in the future. “A signal is still needed for pedestrians and bicyclists,” Hajibabai explains.
The cost of such a system, which would require wide-scale adaptation and replacement of ubiquitous traffic infrastructure to add a fourth light, would be high. As an alternate solution, “a different signal indication could be used, and the traffic light does not have to be white. Signals such as a combination of flashing green and yellow could be used,” Hajbabaie says. “If we go through this route, traffic lights need to be reprogrammed and not replaced, which is significantly less costly.”
This type of alternative could be rolled out sooner, perhaps even in the near future in areas such as ports, as automation in the trucking industry develops.
The next step for the researchers is to test their concept in multiple physical tests in a small-scale AV test bed that they have already developed; after that, full-scale AV test bed simulations will begin.
Catherine A. Cardno, Ph.D., is the editor in chief of Civil Engineering Online.
This article first appeared in Civil Engineering Online as “Proposed ‘white lights’ would assist self-driving.” It appeared in the May/June 2023 print issue of Civil Engineering as “White Lights.”